re-implement result on RESIDE

Question

re-implement result on RESIDE

PhanVinhLong opened this issue 4 years ago · comments

I used your code to train and test on RESIDE V0 dataset but I got the result 0.84/20.57 (SSIM/PSNR) for indoor task. Was I wrong at any step?

Qiming Hu · Answer 1 · Tue Jun 23 2020 12:07:01 GMT+0800 (China Standard Time)

Same here. Bad results.

Qiming Hu · Answer 2 · Fri Jun 26 2020 10:19:56 GMT+0800 (China Standard Time)

@PhanVinhLong Have you found any solutions? plz tell me :)

Aditya Arora · Answer 3 · Wed Jul 08 2020 04:57:24 GMT+0800 (China Standard Time)

Hi
I am also unable to replicate the result for ITS dataset using the command given in README.
Is someone able to get the results by training the model from scratch.
Thanks

IYuan505 · Answer 4 · Mon Jul 20 2020 15:47:26 GMT+0800 (China Standard Time)

@binwang777
I believe so.

小乐 · Answer 5 · Thu Aug 06 2020 20:17:16 GMT+0800 (China Standard Time)

@PhanVinhLong @Rico1214 My result is the same（PSNR≈21 SSIM≈0.87）. Have you solved it？

Qiming Hu · Answer 6 · Thu Aug 13 2020 23:23:56 GMT+0800 (China Standard Time)

@SunLeL The results authors reported are trained on standard, while the data in v0 dataset look bad and unnatural, so we cannot expect ideal results on v0.

Dong-Hwan Jang 장동환 · Answer 7 · Thu May 13 2021 00:21:15 GMT+0800 (China Standard Time)

@kimosalty did you successfully reproduce the results with RESIDE-standard? It seems like even in the RESIDE-standard setting, it doesn't work